A new LP-bound in multivariate Lipshcitz optimization and systems of inequalities
A new method is presented for the unconstrained and linearly constrained multivariate global optimization of a Lipschitz function which calls for one function evaluation and one small LP in each iteration. Numerical aspects and applications are emphasized. In the unconstrained case (which is the most often studied one) a brief introduction is given to the many existing approaches, and it is shown that the new method alleviates different disadvantages of previous methods. In the linear constrained case, in contrast to almost all previous methods, the new approach can be implemented in a practical sense. Applications stress mixture identification problems (which lead to systems of quadratic inequalities over simplices). In this case previous approaches are outperformed. The talk concludes with pointing out the limits of pure Lipschitz optimization and directions of further research.
- OSTI ID:
- 36143
- Report Number(s):
- CONF-9408161-; TRN: 94:009753-0418
- Resource Relation:
- Conference: 15. international symposium on mathematical programming, Ann Arbor, MI (United States), 15-19 Aug 1994; Other Information: PBD: 1994; Related Information: Is Part Of Mathematical programming: State of the art 1994; Birge, J.R.; Murty, K.G. [eds.]; PB: 312 p.
- Country of Publication:
- United States
- Language:
- English
Similar Records
Augmented penality algorithms
Data Locality Enhancement of Dynamic Simulations for Exascale Computing (Final Report)